Natural selection and correlated landscapes of diversity in the great apes
Authors: Murillo Rodrigues 1; Peter Ralph 1,2,3; Andrew Kern 1,3
Affiliations: 1) Institute of Ecology and Evolution, Department of Biology, University of Oregon; 2) Department of Mathematics, University of Oregon; 3) contributed equally
Keywords: Comparative genomics & genome evolution
Levels of genetic diversity along chromosomes – or landscapes of diversity – are correlated between related groups of species, such as flycatchers, monkeyflowers, and aspens. Even though these correlations should decay with divergence time under neutrality, empirical data demonstrate strong correlations persisting over long evolutionary timescales. What could be maintaining strong correlations in landscapes of diversity across pairs species? Natural selection is known to couple genomic features, such as recombination rate and density of functional sites, with levels of diversity. These features are expected to be largely shared between related taxa, so selection and its linked effects could in principle maintain correlations over longer periods of time. Using the great apes as a model system, we investigated whether landscapes of diversity are correlated in this group. Using forward-in-time simulations of the great apes’ evolutionary history under different selective regimes (neutral, background selection, sweeps), we explore how natural selection shapes correlations over time. Further, we test whether variation in mutation rate along chromosomes can generate patterns like the data. We found high correlations among the great apes’ landscapes of diversity, even between pairs of species that diverged over 10Mya. Forward-in-time simulations of the great apes’ history show that the observed correlations far exceed what can be generated under a neutral null model. In simulations with selection and a realistic distribution of functional sites, we found that correlations between landscapes of diversity might be maintained over longer periods of time. Future work includes simulation-based inference of selection parameters in our model.